Bolsonaro Avatar Tests Brazil's Election AI Rules

The Workers' Party and the Brazil of Hope federation filed complaints with Brazil's Supreme Court and Superior Electoral Court, known as the TSE, over the video, according to Valor International. Their filing argued that the appearance violated an order by Justice Alexandre de Moraes barring Jair Bolsonaro from disseminating political or electoral messages directly or through third parties, regardless of the method used.

Reuters reported that leftist and center-left parties also challenged the video in electoral court. The central legal question is not limited to whether viewers could identify the content as synthetic. It also concerns whether an AI representation can communicate a prohibited political message on behalf of a person subject to a judicial speech restriction.

Flavio Bolsonaro's lawyers defended the video before the TSE, according to Valor International. They argued that it clearly identified its use of AI, did not contain an explicit request for votes, and did not qualify as a prohibited deepfake. Reuters reported that the defense also argued the content was harmless and did not combine digitally created images with real ones.

The TSE's March 3 overview of its 2026 election rules says synthetic campaign material must comply with labeling and other restrictions. The official guidance also bars newly created synthetic content that changes the image, voice, or expression of candidates or public figures during the 72 hours before and 24 hours after voting, even when labeled. That final-window rule does not resolve the separate dispute over the July 25 video.

Reuters reported that TSE President Kassio Nunes Marques said AI campaigning is permissible unless it is used to harm someone, while declining to comment on the pending case. A ruling or further guidance could clarify how the court applies its rules when synthetic media is disclosed but carries political speech associated with a restricted individual.

Most election-related AI policy debates have focused on deceptive impersonation: false audio or video intended to make voters believe a candidate said or did something they did not. This case presents a narrower but consequential issue: a synthetic video that announces its artificial origin while reproducing the identity, voice, and political appeal of a real person.

For AI governance practitioners, the dispute illustrates why disclosure requirements and deepfake prohibitions address different risks. Clear labeling can reduce deception about a video's provenance, but it does not by itself resolve questions involving political authorization, campaign rules, judicial orders, or the use of a person's likeness as a proxy for restricted speech.

The immediate case remains specific to the TSE's interpretation of Brazilian election rules and the restrictions imposed on Jair Bolsonaro. It should not be treated as settled law until the relevant courts rule.